tokenize_text: better matching for initials
yohanboniface opened this issue · comments
Yohan Boniface commented
We are talking about textutils.tokenize_text function.
Regarding initials, we have to see if we can do better than now, because it matches also non initials. For example:
... avec des taux en net repli, malgré la perte du triple A. Les montants levés...
Sounds hard to fix :/